T cells Function and Similarity- predictions from statistical and network analysis of the immune repertoire

The Adaptive Immune System is comprised of immune cells that react specifically to different threats. Each T cell has its own T Cell Receptor (TCR) which can bind different target antigens and mount a specific immune response. TCRs have conserved regions but also highly variable ones, most notably the CDR3 region, known to be the major contribution to the binding affinity TCRs have to different antigens. An open question is what the exact relationship of this highly variable sequence and the function of the T cell, both on the single T cell and the full T cell repertoire levels. This relationship should manifest when looking at reactive samples, where the repertoire reacted to a perturbation such as infection or vaccination.

I will present a novel approach to study the connection between sequence similarity and functional similarity, based on data of reactive T cells from mice and patients. I will use analysis tools from statistical physics and graph theory, by formulating the immune repertoire as a connected graph where the TCRs are the nodes, and the edge weights correspond to the similarity. Statistical measures such as Transitivity or the Laplacian matrix eigenvalues learned from the network structure enable us to characterize the immune repertoire and correlate it with the clinical status when the sample was taken. Eventually, this similarity analysis would empower T cell analysis by drawing more statistical power from related TCRs, making new predictions and diagnostic procedures possible.